Nevin Manimala

Med Phys. 2021 Apr 27. doi: 10.1002/mp.14902. Online ahead of print.

ABSTRACT

PURPOSE: Despite the widespread availability of in-treatment room cone beam computed tomography (CBCT) imaging, due to the lack of reliable segmentation methods, CBCT is only used for gross set up corrections in lung radiotherapies. Accurate and reliable auto-segmentation tools could potentiate volumetric response assessment and geometry-guided adaptive radiation therapies. Therefore, we developed a new deep learning CBCT lung tumor segmentation method.

METHODS: The key idea of our approach called cross modality educed distillation (CMEDL) is to use magnetic resonance imaging (MRI) to guide a CBCT segmentation network training to extract more informative features during training. We accomplish this by training an end-to-end network comprised of unpaired domain adaptation (UDA) and cross-domain segmentation distillation networks (SDN) using unpaired CBCT and MRI datasets. UDA approach uses CBCT and MRI that are not aligned and may arise from different sets of patients. The UDA network synthesizes pseudo MRI from CBCT images. The SDN consists of teacher MRI and student CBCT segmentation networks. Feature distillation regularizes the student network to extract CBCT features that match the statistical distribution of MRI features extracted by the teacher network and obtain better differentiation of tumor from background. The UDA network was implemented with a cycleGAN improved with contextual losses separately on Unet and dense fully convolutional segmentation networks (DenseFCN). Performance comparisons were done against CBCT only using 2D and 3D networks. We also compared against an alternative framework that used UDA with MR segmentation network, whereby segmentation was done on the synthesized pseudo MRI representation. All networks were trained with 216 weekly CBCTs and 82 T2-weighted turbo spin echo MRI acquired from different patient cohorts. Validation was done on 20 weekly CBCTs from patients not used in training. Independent testing was done on 38 weekly CBCTs from patients not used in training or validation. Segmentation accuracy was measured using surface Dice similarity coefficient (SDSC) and Hausdroff distance at 95th percentile (HD95) metrics.

RESULTS: The CMEDL approach significantly improved (p < 0.001) the accuracy of both Unet (SDSC of 0.83 ± 0.08; HD95 of 7.69 ± 7.86mm) and DenseFCN (SDSC of 0.75 ± 0.13; HD95 of 11.42 ± 9.87mm) over CBCT only 2DUnet (SDSC of 0.69 ± 0.11; HD95 of 21.70 ± 16.34mm), 3D Unet (SDSC of 0.72 ± 0.20; HD95 15.01 ± 12.98mm), and DenseFCN (SDSC of 0.66 ± 0.15; HD95 of 22.15 ± 17.19mm) networks. The alternate framework using UDA with the MRI network was also more accurate than the CBCT only methods but less accurate the CMEDL approach. Limitation includes analysis on a modest sized dataset.

CONCLUSIONS: Our results demonstrate feasibility of the introduced CMEDL approach to produce reasonably accurate lung cancer segmentation from CBCT images. Further validation on larger datasets is necessary for clinical translation.

PMID:33905558 | DOI:10.1002/mp.14902

J Dent Educ. 2021 Apr 27. doi: 10.1002/jdd.12622. Online ahead of print.

ABSTRACT

PURPOSE: To examine the integration of social determinants of health (SDH) in the US Advanced Education in General Dentistry (AEGD) and General Practice Residency (GPR) programs.

METHODS: This study used an explanatory sequential mixed-methods approach. A 46-question survey was sent to all 265 AEGD and GPR programs in February 2019. Descriptive statistics and multivariate analyses were conducted to identify factors influencing SDH curricular inclusion. A convenience sample of program directors (PDs) was interviewed between June and December 2019. Through content analysis, themes and subthemes were identified.

RESULTS: Of the 265 AEGD and GPR PDs, 111 completed the survey (42% response rate). Almost three-quarters of PDs (72%) agreed that it was important for residents to understand basic SDH concepts. However, programs lacked eight of the 10 surveyed SDH subtopics. The odds of teaching five or more SDH subtopics were 0.09 (95% CI: 0.02-0.41) for programs with none-to-minimal levels of SDH integration in their clinical settings compared to close-to-fully integrated ones. Coding of PD interviews (N = 13) identified five major themes: 1. influences to integrate SDH, 2. training strategies, outcomes, and outputs, 3. reasons for training strategies, 4. barriers and solutions, and 5. future integration goals. Most PDs cited delivering SDH content during patient care and reported time and organizational culture being barriers to more curricular inclusion.

CONCLUSIONS: AEGD and GPR curricula are deficient in SDH content and risk underpreparing residents for caring for the underserved. PDs and organizational leaders must prioritize SDH inclusion in order to train dentists for integrated person-centered care.

PMID:33905531 | DOI:10.1002/jdd.12622

Cancer. 2021 Apr 27. doi: 10.1002/cncr.33515. Online ahead of print.

ABSTRACT

BACKGROUND: Since 2011, the therapeutic landscape of melanoma has changed dramatically because of the adoption of immune checkpoint inhibitor and targeted therapies. The authors sought to quantify the effects of these changes on short-term treatment costs by comparing the first-year cancer-attributable costs in novel (2011-2015) and historical (2004-2010) treatment eras.

METHODS: The authors estimated the first-year cancer-attributable and out-of-pocket (OOP) costs by cancer stage at diagnosis by using a case-control approach. Patients aged ≥67 years with melanoma results were used to calculate the total direct costs of treatment during the first year after the diagnosis of melanoma in the US Medicare population older than 65 years. Costs were reported in 2018 dollars.

RESULTS: Costs increased with the stage at diagnosis. Average first-year cancer-attributable costs per patient for stage IV patients increased significantly by 61.7% from $45,952 to $74,297 after the adoption of novel treatments. Per-patient OOP responsibility decreased by almost 30.8% across all stages of cancer but increased by 16.5% for stage IV patients from 2004 ($7646) to 2015 ($8911). The total direct cost of treatment for persons with melanoma older than 65 years increased by $16.03 million (4.93%) from $324.68 million in 2010 to $340.71 million in 2015. The largest increase in yearly total cost, $23.64 million (56.53%), was observed among stage IV patients.

CONCLUSIONS: The direct cost of melanoma increased significantly in the Medicare population, particularly for advanced-stage disease. Prevention and early detection initiatives may reduce the economic burden of melanoma.

PMID:33905529 | DOI:10.1002/cncr.33515

Pain Med. 2021 Apr 27:pnab154. doi: 10.1093/pm/pnab154. Online ahead of print.

ABSTRACT

OBJECTIVE: Musculoskeletal (MSK) pain conditions are a leading cause of pain and disability internationally and a common reason to seek health care. Accurate prediction of recurrence for health care seeking due to MSK conditions could allow for better tailoring of treatment. The aim of this project was to characterize patterns of recurrent physical therapy seeking for MSK pain conditions and to develop a preliminary prediction model to identify those at increased risk for recurrent care seeking.

DESIGN: Retrospective cohort.

SETTING: Ambulatory care.

SUBJECTS: Patients (n = 578,461) seeking outpatient physical therapy (United States).

METHODS: Potential predictor variables were extracted from the electronic medical record and patients were placed into three different recurrent care categories. Logistic regression models identified individual predictors of recurrent care seeking and Least Absolute Shrinkage and Selection Operator (LASSO) developed multivariate prediction models.

RESULTS: Accuracy of models for different definitions of recurrent care ranged from 0.59 – 0.64 (c-statistic) and individual predictors were identified from multivariate models. Predictors of increased risk for recurrent care included: worker’s compensation and Medicare insurance, comorbid arthritis, post-operative at time of first episode, age range from 44 -64 years, and reporting night sweats/night pain. Predictors of decreased risk for recurrent care included: lumbar pain, chronic injury, neck pain, pregnancy, age range from 25-44 years, and smoking.

CONCLUSION: This analysis identified a preliminary predictive model for recurrence of care seeking of physical therapy, but model accuracy needs to improve to better guide clinical decision making.

PMID:33905514 | DOI:10.1093/pm/pnab154

G3 (Bethesda). 2021 Apr 26:jkab133. doi: 10.1093/g3journal/jkab133. Online ahead of print.

ABSTRACT

This work represents a novel mechanistic approach to simulate and study genomic networks with accompanying regulatory interactions and complex mechanisms of quantitative trait formation. The approach implemented in MeSCoT software is conceptually based on the omnigenic genetic model of quantitative (complex) trait, and closely imitates the basic in vivo mechanisms of quantitative trait realization. The software provides a framework to study molecular mechanisms of gene-by-gene and gene-by-environment interactions underlying quantitative trait’s realization and allows detailed mechanistic studies of impact of genetic and phenotypic variance on gene regulation. MeSCoT performs a detailed simulation of genes’ regulatory interactions for variable genomic architectures, and generates complete set of transcriptional and translational data together with simulated quantitative trait values. Such data provide opportunities to study, for example, verification of novel statistical methods aiming to integrate intermediate phenotypes together with final phenotype in quantitative genetic analyses, or to investigate novel approaches for exploiting gene-by-gene and gene-by-environment interactions.

PMID:33905502 | DOI:10.1093/g3journal/jkab133

Mol Biol Evol. 2021 Apr 27:msab119. doi: 10.1093/molbev/msab119. Online ahead of print.

ABSTRACT

Convergent phenotypic evolution provides some of the strongest evidence for adaptation. However, the extent to which recurrent phenotypic adaptation has arisen via parallelism at the molecular level remains unresolved, as does the evolutionary origin of alleles underlying such adaptation. Here, we investigate genetic mechanisms of convergent highland adaptation in maize landrace populations and evaluate the genetic sources of recurrently selected alleles. Population branch excess statistics reveal substantial evidence of parallel adaptation at the level of individual SNPs, genes and pathways in four independent highland maize populations. The majority of convergently selected SNPs originated via migration from a single population, most likely in the Mesoamerican highlands, while standing variation introduced by ancient gene flow was also a contributor. Polygenic adaptation analyses of quantitative traits reveal that alleles affecting flowering time are significantly associated with elevation, indicating the flowering time pathway was targeted by highland adaptation. In addition, repeatedly selected genes were significantly enriched in the flowering time pathway, indicating their significance in adapting to highland conditions. Overall, our study system represents a promising model to study convergent evolution in plants with potential applications to crop adaptation across environmental gradients. Keyword: convergent adaptation, flowering time, polygenic adaptation, population branch statistic.

PMID:33905497 | DOI:10.1093/molbev/msab119

Clin Infect Dis. 2021 Apr 27:ciab370. doi: 10.1093/cid/ciab370. Online ahead of print.

ABSTRACT

BACKGROUND: Lyme neuroborreliosis peripheral facial palsy (LNB PFP) and idiopathic peripheral facial palsy, Bell’s palsy (BP), are the most common causes of facial palsy in borrelia-endemic areas and are clinically similar. Early treatment with corticosteroids has been shown to be effective in Bell’s palsy and antibiotics improve outcome in LNB, but there is a lack of knowledge on how the addition of corticosteroids to standard antibiotic treatment affects outcome in LNB PFP.

METHODS: This prospective open trial with historical controls was conducted at two large hospitals in western Sweden between 2011 and 2018. Adults presenting with LNB PFP were included in the study group and were treated with oral doxycycline 200 mg b.i.d. for 10 days and prednisolone 60 mg o.d. for 5 days, then tapered over 5 days. The historical controls were adult patients with LNB PFP included in previous studies and treated with oral doxycycline. Both groups underwent a follow-up lumbar puncture and were followed until complete recovery or for 12 months.

RESULTS: Fifty-seven patients were included, 27 in the study group and 30 in the control group. Two patients (6%) in the study group and 6 patients (20%) in the control group suffered from sequelae at end follow up. There was no statistically significant difference between the groups, neither in the proportion of patients with sequelae, nor in the decline in CSF mononuclear cell count.

CONCLUSIONS: Adjunctive corticosteroids neither improve nor impair the outcome for patients with Lyme neuroborreliosis peripheral facial palsy treated with doxycycline.

PMID:33905494 | DOI:10.1093/cid/ciab370

Clin Infect Dis. 2021 Apr 27:ciab190. doi: 10.1093/cid/ciab190. Online ahead of print.

ABSTRACT

BACKGROUND: The gambiense human African trypanosomiasis (gHAT) elimination programme in the Democratic Republic of Congo (DRC) routinely collects case data through passive surveillance and active screening, with several regions reporting no cases for several years, despite being endemic in the early 2000s.

METHODS: We use mathematical models fitted to longitudinal data to estimate the probability that selected administrative regions have already achieved elimination of transmission (EOT) of gHAT. We examine the impact of active screening coverage on the certainty of model estimates for transmission and therefore the role of screening in the measurement of EOT.

RESULTS: In three example health zones of Sud-Ubangi province we find there is a moderate (>40%) probability that EOT has been achieved by 2018, based on 2000-2016 data. Budjala and Mbaya reported zero cases during 2017-18 and this further increases our respective estimates to 99.9% and 99.6% (Model S); and to 87.3% and 92.1% (Model W). Bominenge had recent case reporting, however if zero cases were found in 2021 it would substantially raise our certainty that EOT has been met there (99.0% for Model S and 88.5% for Model W), and this could be higher with 50% coverage screening that year (99.1% for Model S and 94.0% for Model W).

CONCLUSIONS: We demonstrate how routine surveillance data coupled with mechanistic modelling can estimate the likelihood that EOT has already been achieved. Such quantitative assessment will become increasingly important for measuring local achievement of EOT as 2030 approaches.

PMID:33905480 | DOI:10.1093/cid/ciab190

Clin Infect Dis. 2021 Apr 27:ciab192. doi: 10.1093/cid/ciab192. Online ahead of print.

ABSTRACT

Maps of the geographical variation in prevalence play an important role in large-scale programmes for the control of Neglected Tropical Diseases. Pre-control mapping is needed to establish the appropriate control intervention in each area of the country in question. Mapping is also needed post-intervention to measure the success of control efforts. In the absence of comprehensive disease registries, mapping efforts can be informed by two kinds of data: empirical estimates of local prevalence obtained by testing individuals from a sample of communities within the geographical region of interest; digital images of environmental factors that are predictive of local prevalence. In this paper, we focus on the design and analysis of impact surveys, i.e. prevalence surveys that are conducted post-intervention with the aim of informing decisions on what further intervention, if any, is needed to achieve elimination of the disease as a public health problem. We show that geospatial statistical methods enable prevalence surveys to be designed and analysed as efficiently as possible so as to make best use of hard-won field data. We use three case-studies based on data from soil-transmitted helminth impact surveys in Kenya, Sierra Leone and Zimbabwe to compare the predictive performance of model-based geostatistics with methods described in current World Health Organisation guidelines. In all three cases, we find that model-based geostatistics substantially outperforms the current WHO guidelines, delivering improved precision for reduced field-sampling effort. We argue from experience that similar improvements will hold for prevalence mapping of other Neglected Tropical Diseases.

PMID:33905476 | DOI:10.1093/cid/ciab192

PLoS Pathog. 2021 Apr 27;17(4):e1009536. doi: 10.1371/journal.ppat.1009536. Online ahead of print.

ABSTRACT

Skin mononuclear phagocytes (MNPs) provide the first interactions of invading viruses with the immune system. In addition to Langerhans cells (LCs), we recently described a second epidermal MNP population, Epi-cDC2s, in human anogenital epidermis that is closely related to dermal conventional dendritic cells type 2 (cDC2) and can be preferentially infected by HIV. Here we show that in epidermal explants topically infected with herpes simplex virus (HSV-1), both LCs and Epi-cDC2s interact with HSV-1 particles and infected keratinocytes. Isolated Epi-cDC2s support higher levels of infection than LCs in vitro, inhibited by acyclovir, but both MNP subtypes express similar levels of the HSV entry receptors nectin-1 and HVEM, and show similar levels of initial uptake. Using inhibitors of endosomal acidification, actin and cholesterol, we found that HSV-1 utilises different entry pathways in each cell type. HSV-1 predominantly infects LCs, and monocyte-derived DCs, via a pH-dependent pathway. In contrast, Epi-cDC2s are mainly infected via a pH-independent pathway which may contribute to the enhanced infection of Epi-cDC2s. Both cells underwent apoptosis suggesting that Epi-cDC2s may follow the same dermal migration and uptake by dermal DCs that we have previously shown for LCs. Thus, we hypothesize that the uptake of HSV and infection of Epi-cDC2s will stimulate immune responses via a different pathway to LCs, which in future may help guide HSV vaccine development and adjuvant targeting.

PMID:33905459 | DOI:10.1371/journal.ppat.1009536